With the dramatic growth of the image data on the web, there is an increasing demand of the algorithms capable of understanding the visual information automatically.

This thesis mainly focuses on multimodal understanding and Visual Question Answering (VQA) via deep learning methods. For technical contributions, this thesis first focuses on improving multimodal fusion schemes via multi-stage vision-language interactions.

This thesis mainly focuses on cross-modal retrieval and single-modal image retrieval via deep learning methods, i.e. by using deep convolutional neural networks.

In 2018, the number of mobile phone users will reach about 4.9 billion. Assuming an average of 5 photos taken per day using the built-in cameras would result in about 9 trillion photos annually.

Tomography is a powerful technique to non-destructively determine the interior structure of an object.Usually, a series of projection images (e.g.\ X-ray images) is acquired from a range of different positions.

The invention of neural networks marks a critical milestone in the pursuit of true artificial intelligence. Despite their impressive performance on various tasks, these networks face limitations in learning efficiently as they are often trained from scratch.

We are living in an information era where the amount of image and video data increases exponentially.

Generally the data from remote sensing surveys - the scanning of the earth by satellite or aircraft in order to obtain information about it - is screened manually in archaeology. However, constant monitoring of the earth's surface causes a huge influx of data of high complexity and high quality. To…

Vietnam’s record of expanding access to education, and especially its performance on international assessments such as PISA, has raised questions about what Vietnam got right, how, and why and what insights Vietnam’s experiences might offer for efforts at improving the performance of education systems…

In deep reinforcement learning, searching and learning techniques are two important components. They can be used independently and in combination to deal with different problems in AI, and have achieved impressive results in game playing and robotics. These results have inspired research into artificial…

Implementation of the new sound synthesis and analysis method in music composition.

With the emergence of online social networks (OSNs), the way people create and share information has changed, which becomes faster and broader than traditional social media.

A computer can look at, and learn from, many more images than a human specialist. AI systems are rapidly becoming indispensable for medical and biological applications. But they still have to learn how to explain their decisions.

Sewer pipes are an essential infrastructure in modern society and their proper operation is important for public health. To keep sewer pipes operational as much as possible, periodical inspections for defects are performed.

In this thesis we explore machine and deep learning approaches that address keychallenges in high dimensional problem areas and also in improving accuracy in wellknown problems. In high dimensional contexts, we have focused on computational fluid dynamics (CFD) simulations.

Bringing the theoretical world of mathematics and computer science to more applied research areas. That is what Daan Pelt, assistant professor at the Leiden Institute of Advanced Computer Science (LIACS) is trying to achieve with the methods he is developing as a solution to challenges in image proc…

Computers are capable of making incredibly accurate predictions on the basis of machine learning. In other words, these computers can learn without intervention once they have been pre-programmed by humans. At LIACS, we explore and push the borders of what a revolutionary new generation of algorithms…

Radiography is an important technique to inspect objects, with applications in airports and hospitals. X-ray imaging is also essential in industry, for instance in food safety checks for the presence of foreign objects.

The Reinforcement Learning lab conducts research into Reinforcement Learning and Intelligent Combinatorial Algorithms.

Lecture

Simulating human language understanding on the computer is a great challenge. A way to approach it is to represent natural language meanings in logic, and to use logical provers to determine what does and does not follow from a text. What logic is best to use and how natural language meanings are best…

This study aims to provide a long time perspective of human landscape manipulation. Studying the roles of prehistoric foragers in past ecosystems is of great importance to establish the character of past 'natural' landscapes and to enhance the management of current ones.

This project studies the roles of prehistoric foragers in past ecosystems to establish the character of past “natural” landscapes and enhance the management of current ones.

- Which metabolites could be components of NADES? - How can we prove the presence of NADES in nature? - What are the roles of NADES in nature? - How to apply NADES in life sciences?

On dinsdag 26 juni 2018, Linlin Sun defended her doctoral thesis ‘International Environmental Obligations and Liabilities in Deep Seabed Mining’. The doctoral research was supervised by Prof. dr. N.J. Schrijver en Prof. dr. E.C.P.D.C. De Brabandere.

Over several decades, a variety of computational methods for drug discovery have been proposed and applied in practice. With the accumulation of data and the development of machine learning methods, computational drug design methods have gradually shifted to a new paradigm, i.e. deep learning methods…

The recently introduced nature-originated deep eutectic solvents, so-called natural deep eutectic solvents (NADES) are considered as truly green solvents, which composed of natural ingredients found abundantly in organisms.

PhD defence

Thanks to deep learning, computer program AlphaZero plays better chess than the best person. The computer wins because it can predict further ahead than a human being through pattern recognition. Aske Plaat discusses this special development in Dutch newspaper NRC. He expects the major tech companies…

PhD defence

How did Google’s computer programme AlphaGo become so good at board game Go that it could defeat the world champion? On June 29, developer Aja Huang will speak about this during a lecture in the Gorlaeus building.

Building a Deep Cross-media Transfer Network to extract visual relations that relieve the problem of insufficient training data for visual tasks.

How did Google's computer programme AlphaGo become so powerful? On June 29, developer Aja Huang elaborated on this during a lecture in the Gorlaeus building.

A new microtonal approach of the bass clarinet, to further develope the instrument’s capability to produce not only exact quartertones, but also smaller units.

PhD defence

PhD defence

Supercomputer ALICE has been expanded, making it even more powerful and faster. This means researchers and students can work with heavier models. From language research to brain activity scans, this expansion is good news for a variety of disciplines. So who gets to work with this prima donna?

In design optimization problems, engineers typically handcraft design representations based on personal expertise, which leaves a fingerprint of the user experience in the optimization data. Thus, learning this notion of experience as transferrable design features has potential to improve the performance…

Computers are capable of making incredibly accurate predictions on the basis of machine learning. In other words, these computers can learn without intervention once they have been pre-programmed by humans. At LIACS, we explore and push the borders of what a revolutionary new generation of algorithms…

PhD defence

Developing new drugs is a difficult process. With the aid of artificial intelligence, Pharmaceutical scientist Xuhan Liu has developed methods that can help make drug design cheaper and faster. PhD defence on 15 February.